Quick Recap
Endocrine & Metabolic System, Protocol 5/9. Companion to Myxedema Coma (prior protocol) — shares the principle that adrenal insufficiency can coexist with/be unmasked by other endocrine emergencies.
1. Physiology Background
Severe illness/stress normally stimulates the HPA axis (hypothalamic CRH -> pituitary ACTH -> adrenal cortisol) — cortisol production can increase up to 6-FOLD with severe infection, trauma, or burns. During acute illness, corticosteroid-binding globulin (CBG) and albumin levels fall by up to 50%, INCREASING the free/unbound (physiologically active) cortisol fraction — but standard laboratory assays measure only TOTAL cortisol, not the free fraction, creating a fundamental interpretive limitation: a "normal" total cortisol in a hypoalbuminemic critically ill patient may still represent inadequate FREE cortisol, and vice versa. Normal diurnal HPA variation is LOST in critical illness.
2. Definitions — Two Distinct Entities
Chronic adrenal insufficiency (Addison disease): a RARE cause of ICU admission in its own right, distinct from the syndrome below.
Critical Illness-Related Corticosteroid Insufficiency (CIRCI): an inappropriately LOW cortisol INCREASE during acute illness ("relative adrenal insufficiency") — mediated by inflammatory cytokines inhibiting CRH/ACTH release, adrenal cortisol synthesis, AND glucocorticoid receptor translocation/transcription (i.e., tissue-level glucocorticoid RESISTANCE can occur even with adequate circulating cortisol, a third mechanism beyond simple production deficiency). This is the far more common ICU scenario and the primary focus of this protocol's acute management section.
3. Etiology (Primary vs Secondary, and CIRCI-Specific Causes)
Category | Example |
Infection (can cause direct adrenal injury OR precipitate CIRCI) | Sepsis/septic shock, HIV, CMV, toxin-producing S. aureus, fungal disease (histoplasmosis, blastomycosis, cryptococcus), TB |
Medications — suppress CRH/ACTH release | Corticosteroids, megestrol acetate |
Medications — inhibit cortisol synthesis enzymes | Etomidate (even a SINGLE induction dose can transiently suppress adrenal steroidogenesis — relevant given etomidate's frequent use for RSI in critically ill/hemodynamically unstable patients), ketoconazole, metyrapone |
Medications — increase cortisol metabolism | Rifampin, phenytoin |
Malignancy | Adrenal metastases |
Adrenal hemorrhage | Secondary to shock, anticoagulation, meningococcemia (Waterhouse-Friderichsen syndrome), DIC, antiphospholipid syndrome |
Autoimmune | Addison disease |
Hypothalamic/pituitary (secondary AI) | Infection, pituitary tumor/metastases, infiltrative disease (sarcoidosis, histiocytosis), postpartum pituitary necrosis (Sheehan syndrome), trauma (blunt, radiation, surgical) |
Etomidate deserves specific clinical attention: its adrenal-suppressive effect, even from a single RSI dose, has been a subject of real debate regarding its use in septic/critically ill patients specifically — be aware of this when a patient who received etomidate for intubation subsequently develops refractory hypotension.
4. Clinical Presentation
System | Findings |
Cardiovascular | Hypotension usually UNRESPONSIVE to fluid administration, often requiring vasopressors; hyperdynamic hemodynamic profile common; hypotension REFRACTORY to catecholamines; tachycardia UNLESS severe concurrent hypothyroidism is also present |
Metabolic/electrolyte | Hypoglycemia, hyponatremia, hyperkalemia, fever |
Hematologic | Eosinophilia, anemia |
Neuromuscular | Weakness, fatigue, myalgia, arthralgia, headache, memory impairment, depression |
GI | Anorexia, diarrhea, nausea, salt craving, weight loss |
Cutaneous | Vitiligo, alopecia |
In critically ill patients, adrenal insufficiency should be actively SUSPECTED IN ALL PATIENTS with hypotension unresponsive to IV fluids AND vasopressor support — this single clinical trigger (fluid- and vasopressor-refractory hypotension) is the key pattern-recognition cue for this entire protocol, more so than the classic Addison-disease textbook findings, which may be absent or subtle in acute CIRCI.
5. Diagnostic Testing — A Genuinely Unsettled, Evolving Area
Random cortisol: older studies suggested levels >25-34 mcg/dL make relative adrenal insufficiency unlikely; more recent evidence suggests random cortisol has LIMITED diagnostic utility EXCEPT when baseline cortisol is <10-15 mcg/dL (below this, adrenal insufficiency is likely) — do not over-rely on a single random cortisol value in the intermediate range.
Cosyntropin (ACTH) stimulation test: 250 mcg IV cosyntropin, cortisol measured at baseline and 60 minutes. A cortisol INCREASE of <=9 mcg/dL at 60 minutes suggests adrenal insufficiency is LIKELY in critical illness. Measuring cortisol at BOTH 30 and 60 minutes adds NO significant diagnostic value beyond checking at 60 minutes alone — simplify testing accordingly.
Key trial evidence shaping current interpretation:
- Annane et al. 2000 (189 septic shock patients): intermediate/poor prognosis with cosyntropin response <=9 mcg/dL; HIGHEST mortality in patients with BOTH a poor cosyntropin response (<=9) AND a HIGH baseline cortisol (>34 mcg/dL) — this combination (high baseline but poor further stimulation, suggesting the adrenal is already maximally/inappropriately stressed with no reserve) identifies the sickest subgroup
- CORTICUS retrospective arm (Lipiner-Friedman 2007, 477 patients): random cortisol >=15 mcg/dL was NOT an independent predictor of shock reversal/mortality/survival REGARDLESS of cutoff used; but baseline cortisol <15 OR cosyntropin response <=9 mcg/dL predicted longer shock duration and shorter survival
Overall current consensus: adrenal insufficiency is LIKELY when baseline cortisol is <10-15 mcg/dL in critically ill patients; cosyntropin stimulation adds diagnostic value specifically in volume-unresponsive septic shock hypotension, though the cortisol assay's unreliability in critical illness (Section 1) limits how much weight should be placed on any single number.
6. Treatment — Evidence Evolution and Current Recommendation
Given the unreliable nature of cortisol testing in critical illness, the RECOMMENDED approach for refractory septic shock unresponsive to IV fluids and vasopressor support is EMPIRIC TREATMENT rather than test-gated treatment: start hydrocortisone 200-300 mg/day for 5-7 days; the underlying cause of adrenal insufficiency can be investigated AFTER treatment is initiated, not before — do not delay empiric steroids in refractory septic shock pending cosyntropin test results.
Landmark trial evolution (know this trajectory, similar in spirit to the TTM evolution in Post-Cardiac Arrest Syndrome):
- Annane 2002 (300 septic shock patients): hydrocortisone 50mg q6h + fludrocortisone 50mcg daily x7 days vs placebo, AFTER a 250mcg cosyntropin test. In the 229 patients with relative AI (cosyntropin response <=9), corticosteroid treatment showed SIGNIFICANTLY REDUCED MORTALITY and faster vasopressor withdrawal, with NO increase in adverse events. In the 70 patients with a GOOD cosyntropin response (>=9), corticosteroids showed NO benefit — suggesting the test-defined subgroup mattered in this trial.
- CORTICUS (larger, later, methodologically different — randomized up to 72h post-shock-onset rather than within 8h, included sepsis without shock, 11-day course, NO fludrocortisone, and had clinical equipoise/selection bias issues): found NO benefit, including in the cosyntropin-poor-responder subgroup — this discordance with Annane 2002 reflects real methodological differences, not necessarily a true contradiction.
- APROCCHSS 2018 (1,241 patients, same regimen as Annane 2002): showed DECREASED mortality at both 90-day (43.0% vs 49.1%, P=.03) and 180-day (46.6% vs 52.5%, P=.04).
- ADRENAL trial (largest, 3,658 patients, continuous infusion 200mg/day x7 days, NO fludrocortisone): FASTER shock resolution, decreased ICU/ventilator duration, decreased transfusion need — but NO mortality difference at 90-day or 28-day.
Synthesis: newer, larger meta-analyses show MINIMAL OR NO MORTALITY BENEFIT overall, but CONSISTENTLY show improved shock reversal, decreased mechanical ventilation duration, decreased ICU stay, and decreased hospital length of stay — corticosteroid replacement should therefore be provided to ALL patients with refractory septic shock unresponsive to IV fluids and vasopressors, understood as a shock-reversal/morbidity-reduction therapy rather than a proven mortality-reduction therapy.
Caution — not a universal panacea: a case-control study (Britt et al. 2006, burn trauma ICU) found corticosteroid use associated with INCREASED infection rates, increased ICU/ventilator duration, and a TREND toward increased mortality in that specific population — glucocorticoids may NOT be appropriate for all critically ill patients, and this protocol's recommendation is specifically scoped to refractory septic shock, not a blanket ICU steroid policy.
7. Immediate Stabilization (ABCDE) — True Adrenal Crisis
Circulation: aggressive IV fluid resuscitation; empiric hydrocortisone 100mg IV, then 100mg IV q8h (or 200-300mg/day equivalent for refractory septic shock per Section 6) — do not wait for confirmatory labs in a patient with a compelling clinical picture (refractory hypotension + risk factors from Section 3).
Disability: treat concurrent hypoglycemia with dextrose per standard protocol.
Checklist:
8. Drug Dosing Reference
Agent | Dosing | Notes |
Hydrocortisone | 100 mg IV q8h (or 200-300 mg/day continuous infusion x5-7 days for refractory septic shock) | Has both glucocorticoid AND some mineralocorticoid activity |
Dexamethasone | 10 mg IV prior to ACTH stimulation test | Does NOT interfere with the cortisol assay (unlike hydrocortisone), so can be used as a bridge if steroid therapy needs to start before a cosyntropin test is performed |
Fludrocortisone | 50-200 mcg PO daily | Additional mineralocorticoid coverage; used in some trial regimens (Annane 2002, APROCCHSS) but NOT others (CORTICUS, ADRENAL) — role remains somewhat unresolved as a routine addition |
Hydrocortisone toxicities: short-term — hyperglycemia, mood changes, insomnia, GI irritation, increased appetite; long-term — osteoporosis, acne, thin skin, fat redistribution, muscle wasting, cataracts, HPA axis suppression, hypertension, infection risk.
9. Investigations
Baseline cortisol, cosyntropin stimulation test (60-minute value sufficient, 30-minute adds no value) if pursued, electrolytes (Na, K), glucose, CBC (eosinophilia), ACTH level (helps distinguish primary from secondary AI — elevated in primary, low/inappropriately normal in secondary), imaging (adrenal CT if hemorrhage/mass suspected, pituitary MRI if secondary AI suspected) once acute treatment is underway.
10. Organ Support
Aggressive fluid resuscitation; vasopressor support as needed (recognizing catecholamine-refractoriness is itself a diagnostic clue); empiric or test-guided hydrocortisone per Section 6; glucose/electrolyte correction; standard ICU supportive care for the underlying precipitant.
11. Consultation Matrix
Consultation | Trigger | Timing |
Endocrinology | Confirmed/suspected adrenal insufficiency, especially chronic Addison disease workup or complex secondary AI | Within admission |
Infectious Disease | Adrenal hemorrhage/infarction workup (e.g., meningococcemia), infectious precipitant | As needed |
12. Monitoring Framework
Hemodynamic response to fluids/vasopressors/steroids (shock reversal is the primary treatment-response marker per the trial evidence), serial electrolytes and glucose, cortisol/ACTH trend if pursuing formal diagnostic workup, watch for steroid-related complications (hyperglycemia, infection risk) especially with prolonged courses.
13. Complications
Refractory shock/death if unrecognized and untreated, hyperglycemia/infection risk from corticosteroid therapy (Britt et al. caution, especially in burn/trauma populations), missed alternative diagnosis if steroid response is misattributed. Prevention: prompt empiric treatment in refractory septic shock without over-relying on unreliable cortisol assays, judicious (not blanket) use outside the septic shock indication given the burn-ICU counter-evidence. Rescue: standard refractory shock escalation, dexamethasone bridge if testing needs to precede definitive hydrocortisone dosing.
14. Escalation & De-escalation
Escalate: hypotension refractory to fluids + vasopressors -> empiric hydrocortisone without delay.
De-escalate: shock resolving, vasopressors weaned -> continue hydrocortisone course to completion (5-7 days per trial regimens) then taper/discontinue; if chronic adrenal insufficiency confirmed, transition to long-term maintenance replacement (hydrocortisone + fludrocortisone) with endocrinology.
15. ICU Discharge Criteria
Hemodynamically stable off vasopressors, steroid course completed or transitioned to appropriate maintenance dosing if chronic AI confirmed, underlying precipitant identified/treated, electrolytes/glucose stable, endocrinology follow-up arranged for confirmed chronic adrenal insufficiency.
16. Documentation & Medicolegal Checklist
17. Key Guidelines
Annane D, Pastores SM, Rochwerg B, et al. Guidelines for the diagnosis and management of critical illness-related corticosteroid insufficiency (CIRCI): 2017 update. Crit Care Med. 2017;45(12):2078-2088.
18. Landmark Trials
Annane D, et al. 2002 RCT (hydrocortisone + fludrocortisone in cosyntropin-poor-responder septic shock) — mortality benefit. CORTICUS (methodologically distinct, later randomization window) — no benefit. APROCCHSS 2018 — confirmed mortality benefit at 90 and 180 days with the same regimen as Annane 2002. ADRENAL trial (largest, continuous infusion, no fludrocortisone) — improved shock reversal/morbidity but no mortality difference. Britt et al. 2006 (burn trauma ICU) — counter-evidence of harm in a non-septic-shock population.
19. Controversies
The discordance between Annane 2002/APROCCHSS (mortality benefit) and CORTICUS/ADRENAL (no mortality benefit) remains genuinely debated and is attributed to real methodological differences (timing window, population, fludrocortisone inclusion, course duration) rather than a clean resolution — present this honestly rather than picking a side. The role of fludrocortisone as a routine addition to hydrocortisone is unresolved given its inclusion in only some of the positive trials. Cortisol/cosyntropin testing's clinical utility remains limited by fundamental assay issues (measuring total, not free, cortisol; CBG/albumin fluctuation in critical illness), pushing practice toward empiric treatment in refractory shock rather than test-gated treatment. Corticosteroid use outside the septic-shock indication (e.g., burn/trauma ICU) has actual counter-evidence of harm, cautioning against extrapolating the septic shock recommendation too broadly.
20. References
- Nguyen N. Adrenal Insufficiency in Critical Illness. Washington Manual of Critical Care, 4th ed, 2025 (Ch. 30).
- Endocrine Disorders drug dosing reference table. Washington Manual of Critical Care, 4th ed, 2025 (Ch. 95).
- Annane D, Pastores SM, Rochwerg B, et al. Guidelines for the diagnosis and management of CIRCI: 2017 update. Crit Care Med. 2017;45(12):2078-2088.
- Annane D, Sebille V, Charpentier C, et al. Effect of treatment with low doses of hydrocortisone and fludrocortisone on mortality in patients with septic shock. JAMA. 2002;288(7):862-871.
- Sprung CL, Annane D, Keh D, et al. Hydrocortisone therapy for patients with septic shock (CORTICUS). N Engl J Med. 2008;358(2):111-124.
- Annane D, Renault A, Brun-Buisson C, et al. Hydrocortisone plus fludrocortisone for adults with septic shock (APROCCHSS). N Engl J Med. 2018;378(9):809-818.
- Venkatesh B, Finfer S, Cohen J, et al. Adjunctive glucocorticoid therapy in patients with septic shock (ADRENAL). N Engl J Med. 2018;378(9):797-808.
- Lipiner-Friedman D, Sprung CL, Laterre PF, et al. Adrenal function in sepsis: the retrospective CORTICUS cohort study. Crit Care Med. 2007;35(4):1012-1018.
See also: Septic Shock (Cardiovascular System) for the primary clinical context in which this protocol's corticosteroid recommendation applies; Myxedema Coma (Endocrine & Metabolic System) for the reciprocal risk of unmasking adrenal insufficiency during thyroid hormone repletion.